Since non-insulin dependent diabetes mellitus (NIDDM) is a major public health problem, attempts to prevent it are clearly warranted. We hypothesize that a behavioral intervention compared to usual care will: 1) be successful in maintaining modest weight loss; 2) be sustained over a 5 year period; 3) reduce the incidence of NIDDM among persons with impaired glucose tolerance (IGT); and 4) increase the incidence and duration of remissions among persons with screen-detected NIDDM. In addition, 5) persons receiving the intervention will have significant reductions in insulin resistance and associated lipid disorders; and 6) persons receiving the intervention will have significantly less progression of carotid wall thickness than persons in the usual care group. The trial is proposed among Hispanic and non-Hispanic white members of the Kaiser Permanente Health Plan, which has 98,745 members aged 40-69 in the Denver metropolitan area. One hundred each of Hispanic and non-Hispanic white subjects are proposed for enrollment. Screening will be done among members who are 120% of ideal body weight using two OGTT's to identify persons with "stable" impaired glucose tolerance (IGT) and previously undiagnosed NIDDM. Each of these groups must be entered as separate cohorts due to the substantial numbers of differences in the groups. We are proposing a 3 phase behavioral intervention group aimed at modest weight loss, reduction of dietary fat and stepped increases in physical activity tailored to the subject to enhance long-term compliance. A drug arm is not proposed by our center since we believe that there is little published support for the efficacy of any class of pharmaceutical agent at this time. We have suggested designs that we feel will accommodate both behavioral and drug interventions, if both are finally selected. The primary comparison group will be to usual care, randomized after screening and eligibility determination. All groups will receive usual care for other conditions. The primary endpoint is the development of NIDDM by typical symptoms and/or a yearly OGTT on two occasions meeting WHO criteria. Among persons with NIDDM, remission to normal or impaired glucose tolerance is the proposed endpoint. Secondary end points include obesity, clinical and preclinical (B-mode carotid wall thickness) cardiovascular disease, hypertension, lipid, and insulin resistance and secretory measures. Sample size estimates indicate that sufficient power should exist in a trial of approximately 2000 subjects to detect a 25% intervention effect at 80% power if NIDDM incidence is 7% per year. Analysis will be by intention to treat, wit secondary analyses exploring changes in rates by levels of compliance to the intervention. The Kaiser population is large enough to allow identification of more than 200 subjects, should that be required.